Earth system models (ESMs) use photosynthetic capacity, indexed by the maximum Rubisco carboxylation rate (V_cmax), to simulate carbon assimilation with an assumed dependence on leaf nitrogen determined from soil fertility. In contrast, new theory, based on biochemical coordination and co-optimization of carboxylation and water costs for photosynthesis, has found that optimal V_cmaxcan be predicted from climate. Building on this work, we test this new theory on a global, field-measured V_cmaxdataset of >100 within canopy light gradients from a variety of ecosystem types around the world. We first compare model-predicted V_cmaxto observed V_cmaxvalues to examine the interaction of biophysical constraints with light availability. Second, we examine the sensitivity of our V_cmaxpredictions to soil indices as proxies for N and water supply, temperature and other leaf traits strongly correlated to photosynthetic capacity such as leaf nitrogen per area (N_a) and leaf mass per area (LMA). Finally, we review and contrast our findings in the context of other optimality theories based on N supply from the soil driving global photosynthetic capacity.